WxS-QC - a quality control pipeline for human Whole-Genome and Whole Exome sequencing cohorts

Motivation Whole-exome (WES) and whole-genome (WGS) sequencing are rapidly becoming preferred methods for population-scale analysis of the human genetic landscape. However, there are currently no standardized quality control (QC) pipelines for human WES and WGS datasets. Moveover, there are no open datasets that can be used to test QC pipelines, because most projects (like 1000 genomes and gnomAD) publish only post-QC results. Results We present WxS-QC, a powerful, scalable, and convenient pipeline for the QC of human WGS and WES cohorts, developed at the Wellcome Sanger Institute (WSI). Our pipeline is based on a deep refactoring of the gnomAD quality control pipeline code and is aligned with current best practices in WGS/WES cohort data QC. It offers a set of novel QC techniques, automatic export of resulting graphs and summary tables, excellent performance and scalability, and incorporates comprehensive documentation. To test our pipeline and similar solutions, we also assembled an open dataset with all required metadata. Availability and implementation The pipeline code is written in Python using the Hail library and is freely available under the BSD-3 license here: https://github.com/wtsi-hgi/wxs-qc . It can run in any UNIX-like environment and has been able to efficiently process cohorts of up to 200,000 whole-exome samples with the potential to handle bigger datasets, depending on the available hardware. The detailed description of the pipeline, resources and test data is available in the pipeline documentation: https://github.com/wtsi-hgi/wxs-qc/blob/main/README.md The open test dataset is available to download from https://wxs-qc-data.cog.sanger.ac.uk/wxs-qc_public_dataset_v3.tar . An example of test dataset analysis is available in the supplementary materials.